1. The spatial distribution of global Rs sites

In daily and monthly time scale, we used this approach a lot Rs_mat_Rs_diurnal ()
We have much more measurements Rs in mid-latitude regions, where developed countries are mostly located
It is difficult to measure soil respiration all year around in cold regions, but critical because of high rates of climate change and large soil C stocks
Less-developed countries are constrained by lack of resources, and thus we do not have enough measurements from tropical regions (Xu and Shang 2016)

Global spatial distribution of soil respiration sites

Rs data from cold regions are more important, but how to increase the number of measurements? * Make equipment work normally in very cold and remote conditions * Increase resources devoted to Rs measurements * Methodological improvements: for instance, measuring once per day to get daily mean

Rs measured at diurnal soil temperature

Measure once per year to get annual Rs mean? This is Bahn’s approach Bahn et al. (2004) Biogeosciences

Rs measured at mean annual soil temperature

Rs measured at annual mean temperature is ~~linearly~~ related with annual Rs rate
Rs at mean temperature: soil respiration measured at annual mean temperature, monthly mean temperature, and/or daily mean temperature
Rs_annual ~ Rs_mat (“Rs_bahn”)
Rs_annual = 455.8 * Rs_mat^1.0054 (R² = 0.94, p<0.001)

2. The objects of this analysis are

Examine whether Rs measured at annual mean soil temperature can represent annual Rs rate? (We now have an order of magnitude more data than Bahn et al.)
Annual mean air temperature (e.g., average of 12 monthes’ air temperature of 2000) / Mean annual air temperature (1964-2014)
If not, when and what is the mechanism?
Update the model?

3. Methods

Data

Use SRDB_V4 – Rs_Annual
Annual mean soil temperature (reported in the papers or can be calculated with simple assumptions)
Relationship between Rs and soil temperature (SRDB_V4)
Air temperature (University of Daleware global precipitation and air temperature data, 1964-2014, half degree spatial resolusion)
823 records from 253 studies

## Warning: Removed 147 rows containing non-finite values (stat_smooth).

## Warning: Removed 147 rows containing missing values (geom_point).

## Warning: Removed 687 rows containing non-finite values (stat_smooth).

## Warning: Removed 687 rows containing missing values (geom_point).

## Warning: Removed 125 rows containing non-finite values (stat_smooth).

## Warning: Removed 125 rows containing missing values (geom_point).

## Warning: Removed 696 rows containing non-finite values (stat_smooth).

## Warning: Removed 696 rows containing missing values (geom_point).

Statistics

According the the relationship between Rs and Ts, we can estimate Rs_mat based on the annual mean soil temperature, T_Annual, and/or MAT
Use the Bahn Biogeoscience model (Rs_annual = 455.8 * Rs_mat^1.0054) to predict Rs_annual based on Rs_mat
Comparing Rs_annual and Rs_annual_bahn to evaluate the performance of Bahn model across the global

Update Bahn’s model

If Bahn (2010) model does not predict Rs_annual in all conditions
Update Bahn (2010) model (e.g., including drought parameter, other parameters?)
Regression tree modeling?

3.2 test the relationship between Rs_annual and Rs_mat

## Wed Apr  3 17:58:33 2019  -------------------+++++-------------------
## Wed Apr  3 17:58:33 2019  Bahn relationship for these data:
## 
## Call:
## lm(formula = Rs_TAIR ~ Rs_annual, data = sdata)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -1006.45   -84.72    14.90    93.38   964.58 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.64266   13.67584   0.047    0.963    
## Rs_annual    0.85184    0.01466  58.126   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 185.1 on 821 degrees of freedom
## Multiple R-squared:  0.8045, Adjusted R-squared:  0.8043 
## F-statistic:  3379 on 1 and 821 DF,  p-value: < 2.2e-16

## [1] "test intercept=0 and slope=1"
## [1] "p_intercept = 0.9625, p_slope = 0"

3.3 Ts sources (MGRsD, MGRsD_TAIR, From paper, Rs_Ts_relationship)

## Saving 7 x 5 in image

3.4 Annual Rs or Ts coverage effect

## Saving 7 x 5 in image

## Saving 7 x 5 in image

3.4.1 Test whether outliers affect the regression (need update)

Conclusion: need to update the code and results

## Saving 7 x 5 in image

## Saving 7 x 5 in image

3.4.2 Effect of maximum allowed divergence between global climate data set and site-specific air temperature

Does TAIR_dev and TAIR_LT<_dev affect the relationship – YES!!!!!
TAIR_LTM_dev = with( srdb, abs( MAT_Del - MAT ) )
Does TAIR_LTM_dev () pull the slope off 1? – YES!!!!!
TAIR_dev <- with( srdb, abs( TAnnual_Del - Study_temp ) )
Figure E. Effect of maximum allowed divergence between global climate data set and site-specific air temperature, when given. As we throw out data points with high divergence, R2 goes up (top panel) and RSE goes down (bottom, g C m-2 yr-1).

## Saving 7 x 5 in image

## Saving 7 x 5 in image

3.4.3 Effect of maximum allowed divergence between annual precipitation from paper and Del

4. Results

4.1 Using Ts, TAnnual or MAT

4.1.1 Using soil temperature

## Saving 7 x 5 in image

4.1.2 Using T_Annual

## Saving 7 x 5 in image

4.1.3 Using MAT

## Saving 7 x 5 in image

4.2 Analysis when Rs_mat cannot represent Rs_annual

4.2.2 Does Ecosystem_type affect the relationship between Rs_annual and Rs_mat?

## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?

## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

4.2.3 Does Meas_method affect the relationship?

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

## Saving 7 x 5 in image

4.2.4 RA- or RH-dominated effect?

## Saving 7 x 5 in image

4.2.5 Biome effect?

## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?

## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?
## geom_path: Each group consists of only one observation. Do you need to
## adjust the group aesthetic?

4.2.6 TAIR and precipitation variability effect?

## Saving 7 x 5 in image

## Saving 7 x 5 in image

4.2.7 Drought effect?

## Saving 7 x 5 in image

## Saving 7 x 5 in image

5. Discussion & questions

6. More analysis in the future

1 Using SD information with boosting?
2 Use Rs_mat predict Rh?
3 Use this approach estimate global Rs
4 Think about application
5 Update bahn model with more predictors or using regression tree method?

Can soil respiration measured at MAT represent annual soil respiration?

Jinshi

March 26, 2019